The pundits and the tipsters never talk about staking. They'll give you a tip till they're blue in the face, but they won't say how much you should have on.
If they do suggest a stake, they'll refer to points:
1 point each-way at 16/1
2 points to win at 11/4
But how much is a point?
If you say that 1pt = 1/100th of your betting bank, or 1/50th, you start to make some sense of the risks you're managing.
In this blog 1pt = 1% of the bank.
A couple of years ago, on the betfair General Betting forum, I trialled an experiment in what's known as Kelly Staking. In short, Kelly is a mathematical formula to help you arrive at the "correct" or "optimum" stake.
If you decide to back a horse at 6/1 on the basis that it's overpriced, and you're able to say that in your opinion the odds should be 5/1, then Kelly's formula will recommend a stake.
Here's a Kelly stake calculator: kelly calculator
 |
| John Kelly: not a horseplayer |
Thankfully it was a winning experiment (in spite of it including a long losing streak of 24 bets), but the point of the exercise was to chart the performance of different staking strategies. And Kelly comfortably outstripped two other staking plans, as follows:
Kelly staking made a profit of 109 pts
1% level staking made a profit of 26 pts
Staking to win 10% made a profit of 35 pts
Follow this link for the numbers in detail: kelly on trial
During the experiment, some of the good people of betfair were very down on Kelly, suggesting that it was "too risky". I found that it wasn't, and I'm risk-averse. There were cases which threw up fairly high stakes, but these were few and far between.
If you're a loser, however, Kelly will make you lose more quickly...
To be continued.
No comments:
Post a Comment